Combinatorial optimization model and MIP formulation for the structural analysis of conditional differential-algebraic systems

In this paper we consider the structural analysis problem for differential-algebraic systems with conditional equations. This problem consists, given a conditional differential-algebraic system, in verifying if the system is structurally nonsingular for every state, and if not in finding a state in which the system is structurally singular. We give a formulation for this problem as an integer linear program. This is based on a transformation of the problem into a matching problem in an auxiliary graph. We also show that the linear relaxation of that formulation can be solved in polynomial time. Using this, we develop a Branch-and-Cut algorithm for solving the problem and present some experimental results.     

A deep stacked wavelet auto-encoders to supervised feature extraction to pattern classification

The major issue in pattern classification is in the extraction of features in the training phase. The focus of this work is on combining the ability of wavelet networks and the deep learning techniques to propose a new supervised feature extraction method to pattern classification. This new approach allows the classification of all classes of the dataset by the reconstruction of a Deep Stacked wavelet Auto-Encoder. This Network is obtained after a series of wavelet Auto-Encoders followed by a Softmax classifier at the last layer. Finally, a fine-tuning is applied for the improvement of our result using a back propagation algorithm. Our approach is tested with different image datasets which are the COIL-100, the APTI and the ImageNet datasets and is also tested with two other audio corpuses that contain Arabic words and French words. The experimental test demonstrates the efficiency of our network for image and audio classification compared to other methods.     

An optimization strategy based on a metamodel applied for the prediction of the initial blank shape in a deep drawing process

The transformation of the sheet into a product without failure and excess of material in a deep drawing operation means that the initial blanks should be correctly designed. Therefore, the initial blank design is a critical step in deep drawing design procedure. Consequently, an easy approach for engineers in predicting the initial blank shape is necessary to reduce wastage in material and to overcome the large time consumed in the classical approaches. Thus, the aim of the present investigation is to propose an automatic procedure for the quick sheet metal forming optimization. In fact, a metamodel will be build based on artificial neural networks which will be coupled then with an optimization procedure in order to predict the initial blank shape in a rectangular cup deep drawing operation. The metamodel is built from the finite element simulations using ABAQUS commercial code. This procedure allows a significant reduce of the CPU time compared to classical optimization one. The results show that the desired shape is in good agreement with the one calculated using the optimized blank shape.